AI-first onboarding plan builder
An agent skill that turns a role into a 30-60-90 day onboarding plan that protects the leverage mindset from day one — sandboxes, guilds, and a first-system milestone.
AI-First Onboarding Plan Builder
Purpose
Convert a job description and scorecard into a 30-60-90 day onboarding plan that installs the leverage mindset from day one. Hiring an Architect and then onboarding them like a Manual Pro is how the investment leaks away.
Why this matters
The first 90 days set the default behaviour for years. If the new hire's first tasks are purely manual, manual becomes the norm. If week one gives them a sandbox, a shared prompt library, and a "build one system" milestone, leverage becomes the norm.
Instructions
- Read the job description and scorecard. Extract the core outcome of the role, the top three recurring workflows, and which of the six pillars are most critical.
- Design the first-week sandbox: safe access to the real tools (AI assistant, automation platform, internal data) with a low-stakes task that lets them iterate without breaking production.
- Set a first-system milestone for day 30: the new hire ships one workflow, template, or prompt library that removes a manual step for the team. This is non-negotiable — it is the Architect proof.
- Build the guild hook for day 60: introduce them to the internal "AI Guild" or peer group where prompts and workflows are shared. Give them one thing to contribute and one thing to adopt.
- Define the day-90 leverage review: a 30-minute conversation where they present one system they built, one thing they deliberately did not automate and why, and one failure where AI was trusted too far.
- For each phase, list: the goal, the specific tasks, the success signal, and the person responsible.
Output format
A markdown table grouped by phase:
| Day | Goal | Tasks | Success signal | Owner |
Followed by three "guardrails" — the behaviours the manager must model so the plan holds (e.g. "never hand the new hire a task you have already systematised — point them at the system instead").
Anti-patterns to avoid
- "Read the wiki and shadow calls for two weeks." Zero leverage signal, zero momentum.
- Banning AI during onboarding "until they learn the basics." You teach them to work like a Manual Pro.
- No shared prompt library on day one. Without it, the Multiplier pillar has nowhere to land.
Rules
- The day-30 first-system milestone is mandatory. A plan without it is incomplete.
- Every task should have an observable success signal, not a checklist of activities.
- If the role has no sandbox access in week one, flag it as a blocker — the plan cannot start until it is resolved.
Want this installed in your hiring process?
We run a 30-minute intro call to map your roles against the AI-first framework.
Book a 30-min intro callRelated reading
AI-first CV screening rubric
An agent skill that screens CVs for leverage signals instead of buzzwords — connectors over users, the lazy-AI check, and a structured pass/hold/decline call.
AI-first reference check script
An agent skill that runs a 15-minute reference call probing the six pillars through a former manager, and returns a structured risk read before the offer.
AI-first take-home challenge designer
An agent skill that designs take-home challenges where a perfect result is useless without a process artifact — so you can tell who built it, not who pasted it.